[
https://issues.apache.org/jira/browse/ARROW-12231?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17315721#comment-17315721
]
Weston Pace commented on ARROW-12231:
-------------------------------------
To add to this. The reader-backed dataset is not restartable while a
RecordBatchVector/Table-backed dataset could be restartable, or it could remove
its ownership of the record batches / table as it scans (allowing batches
already scanned to be freed).
> [C++][Dataset] Separate datasets backed by readers from InMemoryDataset
> -----------------------------------------------------------------------
>
> Key: ARROW-12231
> URL: https://issues.apache.org/jira/browse/ARROW-12231
> Project: Apache Arrow
> Issue Type: Improvement
> Components: C++
> Affects Versions: 4.0.0
> Reporter: Weston Pace
> Assignee: David Li
> Priority: Major
> Labels: dataset, datasets
>
> From ARROW-10882/[https://github.com/apache/arrow/pull/9802]
> * Backing an InMemoryDataset with a reader is misleading. Let's split that
> out into a separate class.
> * Dataset scanning can then use an I/O thread for the new class.
> * Longer-term, we should interface with Python's async.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
